Synthetic bone graft material made to closely resemble natural bone would be a useful replacement for natural bone. Acceptable synthetic bone can avoid the problem of availability and harvesting of autogenous bone and the risks and complications associated with allograft bone, such as risks of infection, disease, and viral transmission.
Natural bone is a composite material consisting of both water and organic and inorganic solid phases. Bone has a hard structure because its organic extracellular collagenous matrix is impregnated with inorganic crystals, principally hydroxyapatite (Ca10(PO4)6(OH)2). Calcium and phosphate account for roughly 65% to 70% of the bone's dry weight. Collagen fibers compose approximately 95% of the extracellular matrix and account for 25% to 30% of the dry weight of bone. The organic material gives bone its flexibility and resilience, while the inorganic material gives bone its strength and rigidity (modulus), and the organization of the two phases provides a high degree of toughness to the composite. A thorough review of bone structure from the angstrom level (mineral crystal) to the micron level (lamellae) has been presented (Weiner, S. et al. [1992] FASEB, 6:879-885).
Surrounding the mineralized collagen fibers is a ground substance consisting of protein-polysaccharides, or glycosaminoglycans, primarily in the form of proteoglycan macromolecules. The glycosaminoglycans serve to cement together the various layers of mineralized collagen fibers. The individual collagen molecules self-assemble to form triple helices, which assemble into collagen fibrils, which then assemble into microscopic fibers. Within the packing of the collagen fibrils/fibers are distinct gaps, sometimes called hole zones. These hole zones are created by the staggered arrangement of tropocollagen molecules (triple helical rods), which leads to periodicity of the hole and overlap zones. Various models have been proposed where these hole zones are completely isolated from each other, or are contiguous and together form a groove. Within these hole zones, mineral crystals form. The mineral crystals in final form nucleate and grow within the fibrils (intrafibrillar mineralization), as well as into the interstitial spaces (interfibrillar mineralization) (Landis, W. J. et al. [1993] J. Struc. Biol. 110:39-54). The mineral crystals in final form are a carbonated apatite mineral (dahllite), but initially may form as an amorphous calcium phosphate phase, which then transforms into the apatite (or possibly via an octacalcium phosphate precursor, which naturally forms plates). The apatite platelets of bone are of nanoscopic dimensions (only a few unit cells thick), and are densely packed into the type I collagen fibrils due to the intrafibrillar mineralization mechanism, and are well oriented with their c-axis (in the [001] direction) parallel to the long axis of the collagen fibrils. Because of the nature of the packing, the orientation of the collagen fibrils will determine the orientation of the mineral crystals (Martin, R. B. et al. [1998] “Skeletal Tissue Mechanics”, Springer-Verlag Publishers, New York, N.Y.).
There are numerous biocompatible artificial bone substitutes currently on the market. Of these substitutes, none successfully mimics the composite or microstructure of bone. For example, man-made ceramic composites have some of the desired properties of natural bone (such as matching of modulus), but are notoriously brittle and prone to cracking. By contrast, biological ceramics like bone and teeth resist cracking, with a high toughness and stiffness. It is the nanostructured architecture that leads to mechanical properties that are unique to bone, which are not readily duplicated by polymers (which are not strong or stiff enough), or ceramics (which are brittle and lack toughness, and usually not bioresorbable). These mechanical properties are important because of the body's natural repair processes, in which bone is a living tissue and the cells respond according to the stresses they sense in their surrounding tissue (according to Wolff's Law). If an implant material has too high of a modulus (stiffness), the cells tend to resorb the surrounding bone due to the phenomenon of stress shielding (the stiffer material carries more of the load than the surrounding bone).
A logical choice of materials for a synthetic bone substitute would be a collagen-hydroxyapatite composite; indeed, many have attempted to mineralize collagen in vitro, but the preparation of such a composite has been limited by the ability to achieve the high mineral loading that is attained biologically by intrafibrillar mineralization. An associated periodic contrast pattern is commonly observed by transmission electron microscopy (TEM) of collagen fibers (Carter, J. G. [1990] Skeletal Biomineralization: Patterns, Processes and Evolutionary Trends, Volume 1, Van Nostrand Reinhold Publishers, New York, N.Y.; Hodge, A. J. et al. [1963] “Recent studies with the electron microscope on ordered aggregates of the tropocollagen molecule”, in Aspects of Protein Structure, Ramanchandran, G. N. (ed.), pp. 289-300, Academic Press, London, England; Katz, E. P. et al. [1989] Connect. Tissue Res., 21:49-159). From tomographic imaging of naturally mineralizing turkey tendon (which is considered a model of secondary bone formation), there is evidence that the hydroxyapatite crystals first appear within the hole zones of collagen, and then spread throughout the fibrils, leading to the array of iso-oriented nanocrystals of highly organized hydroxyapatite [HAP] embedded within the organic matrix (Landis, W. J. et al. [1993] Structural Biology, 110:39-54; Landis, W. J. et al. [1991] Connect. Tissue Res., 25:181-196). Alternatively, there has been evidence that the collagen fibers contain an amorphous substance during the early stages of bone formation, referred to by Bonnuci as an “inorganic substance in bands” (ISBs), which then crystallizes into the more commonly observed platy crystals (Bonnuci, E. Calcification in Biological Systems [1992] CRC Press Boca Raton, Fla.).
From a materials engineering perspective, the nanostructure of bone is intriguing and can be difficult to define. For example, it is not clear whether bone is more accurately characterized as a polymer-fiber-reinforced ceramic-matrix composite or a ceramic-nanoparticle-reinforced polymer-matrix composite. The two phases are so intimately linked that the mechanical properties are distinctly different than ceramics or polymers, and therefore are difficult to reproduce. To date, scientists do not have a complete understanding of how bone is formed, even at this most basic level of structure. However, it is likely that the nanostructured architecture plays a role in the toughness of bone. Obviously, cellular control is important in biomineralization, and in the case of bone, helps to build its hierarchical structure (i.e., lamellae and osteons), but even the physicochemical mechanism for generating this nano-architecture has not been elucidated. Because intrafibrillar mineralization does not occur simply by attempting to crystallize collagen in vitro using supersaturated solutions of HAP (crystals only nucleate heterogeneously on the surface of the collagen fibers), it is generally assumed that nucleating proteins must be present within the gaps of the collagen fibrils.
It is understood within the biomineralization community that acidic proteins can act as inhibitors to crystal nucleation or growth (Addadi, L. et al. [1987] Proc. Natl. Acad. Sci. USA, 84:2732-2736; Addadi, L. et al. [1992] Angew. Chem. Int. Ed. Engl. 31:153-169; Mann, S. et al. [1983] Structure and Bonding, 54:125-174; Mann, S. et al. [1989] “Crystallochemical Strategies in Biomineralization” in Biomineralization-Chemical and Biochemical Perspectives. Mann, S., Webb, J., and Williams, R. J. P. (eds.), 33-62 (VCH Publishers, N.Y., N.Y.)). In the case of crystal growth, it has been shown that selective inhibition of growth along stereospecific crystallographic planes can lead to a change in crystal morphology (Addadi, L. et al. [1985] Angew. Chem. Int. Ed. Engl., 24:466-485). In at least a few cases, acidic proteins have been shown to promote crystal nucleation (Addadi, L. et al. [1987] Proc. Natl. Acad. Sci. USA, 84:2732-2736; Greenfield, E. M. et al. [1984] Amer. Zool., 24:925-932). It has also been shown that if the inhibitory action of a macromolecule is not complete, certain conditions lead to the induction (stabilization) of an amorphous liquid-phase precursor (Gower, L. B. et al. [2000] J. Crystal Growth, 210(4):719-734), which can have a profound consequence on crystal morphology since transformation of an amorphous precursor does not proceed via the same mechanism as traditional solution crystal growth (Mann, S. et al. [1989] “Crystallochemical Strategies in Biomineralization” in Biomineralization-Chemical and Biochemical Perspectives. Mann, S., Webb, J., and Williams, R. J. P. (eds.), 33-62 (VCH Publishers, N.Y., N.Y.)). Certain features of this polymer-induced liquid-precursor (PILP) process suggest that this mechanism may occur during morphogenesis of calcium carbonate biominerals in invertebrates (Gower, L. A. [1997] “The Influence of Polyaspartate Additive on the Growth and Morphology of Calcium Carbonate Crystals,” Doctoral Thesis, Department of Polymer Science and Engineering, University of Massachusetts at Amherst, 1-119).
The present inventors have demonstrated that when the PILP phase is deposited onto an amorphous glass substrate, continuous thin films (˜500 nm thick) of CaCO3 are formed (Gower L. A. et al., “Deposition of Calcium Carbonate Films by a Polymer-Induced Liquid-Precursor (PILP) Process” Journal of Crystal Growth, 2000, 210(4):719-734). Organic substrates, such as type-I collagen, have also been used as substrates on which to deposit the PILP phase, resulting in intrafibrillarly mineralized organic/mineral composites (Olszta M. J. et al., [2003] Connective Tissue Research, 44:326-334; Olszta, M. J. et al., [2003] Calcified Tissue International, 72(5): 583-591). Neither of the previous substrates used were of a crystalline nature and, therefore, epitaxial interaction within the crystallization mechanism was not possible.